-
Notifications
You must be signed in to change notification settings - Fork 134
FAQ
How can multiple developers run the same app multiple times in the same account without affecting each other?
You can override the namespace, which is a prefix that mu uses on all CloudFormation stacks (default is mu
). There are 3 ways to set the namespace, in order of precedence:
- Via cli flag:
mu -n foo env ls
- Via environment variable:
MU_NAMESPACE=foo
- Via mu.yml:
namespace: foo
You can provide custom tags in your mu.yml
that gets merged into the templates that mu creates. To learn more, check out Custom CloudFormation.
Yes! For more details, check out the IAM wiki page.
How can I customize the CloudFormation stacks that mu creates for my environment/service/database/pipeline?
You can provide custom CloudFormation YAML in your mu.yml
that gets merged into the templates that mu creates. To learn more, check out Custom CloudFormation.
Yes via custom CloudFormation. Check out the example for enabling case insensitive table names.
Yes via custom CloudFormation. Check out the example for enabling sticky sessions.
Yes via custom CloudFormation. Check out the example for using custom CIDR blocks.
Yes! You can target an existing VPC in your mu.yml
. Check out the vpc-target example to learn more.
Yes! You can provide CloudFormation in your mu.yml
to define the VPN. Check out the vpn example to learn more.
Yes! You can configure your environments to use Fargate inside your mu.yml
. Check out the ecs-fargate example to learn more.
Yes! You can configure your environments to use EKS inside your mu.yml
. Check out the eks example to learn more.
Yes! You can configure your environments to use EC2 and CodeDeploy rather than ECS inside your mu.yml
. Check out the ec2-provider example to learn more.
Why is there still a CloudFormation stack named mu-bucket-codepipeline
after I terminated my pipeline?
The mu-bucket-codepipeline
stack contains an S3 bucket for CodePipeline artifacts. It is shared by all pipelines that are deployed in a given region for an account. Therefore, since it is shared, it isn't safe to assume it can be deleted whenever a pipeline is terminated. Additionally, the bucket is likely not empty which would cause an error when CloudFormation tried to delete it. Alternatively, the bucket could have a DeletionPolicy
of Retain
, but that would cause a conflict the next time pipeline was created and the mu-bucket-codepipeline
stack was recreated.
Therefore, the best approach was to retain the CloudFormation stack that managed the bucket. If you would like to delete the stack, then you should first empty the bucket with aws s3 rm s3://mu-codepipeline-<region>-<account_id> --recursive
and then delete the CloudFormation stack.
*** WARNING *** - this may cause data loss
To remove all the CloudFormation stacks that mu created for your pipelines, environments, databases and services in a given namespace, just run mu purge
.
From the command line, enter mu help
to see the list of available commands.
You can also find more details on the CLI page.
mu (or μ) is the 12th letter of the greek alphabet and is often used to represent "micro", a unit prefix in the metric system. So mu is a reference to a microservice architecture.
mu is pronounced "mew" (:cat2:), not "moo" (:cow2:). Still unsure...listen!
Awesome! Check out the contributing guidelines to get involved.
Ask on Gitter!